How to detect elongated or fibroblast like cells using Cell Detection?

Tumor (1.5 MB)

Sample image and/or code

  • Upload an original image file here directly or share via a link to a file-sharing site (such as Dropbox) – (make sure however that you are allowed to share the image data publicly under the conditions of this forum).
  • Share a minimal working example of your macro code.


The image is a RNA-Scope (5 channels) of a tumor.

  • What is the image about? Provide some background and/or a description of the image. Try to avoid field-specific “jargon”.

Analysis goals

Detect number of spot for each channel per cell.

  • What information are you interested in getting from this image?


Cells are not well detected. Elongated cells are no recognized. How to set the parameters to recognize different cells or how to select them manually or correct the automatic selection.
I used Cell Detection and the Subcellular spot detection using QuPath. How to combine to staining to positively detect the correct shape of the cells?

  • What stops you from proceeding?
  • What have you tried already?
  • Have you found any related forum topics? If so, cross-link them.
  • What software packages and/or plugins have you tried?

As far as I can see, you can’t using basic QuPath methods. You have no cell border markers to delineate edges, and your cells are not spread far enough apart to distinguish individual cells.

There may be options using things like DeepCell and one of your channels (since 3 of your channels are almost exactly the same.

If you decide to proceed, I would definitely recommend better quality images. Fast scan speeds with only 2 averaging are going to hurt your accuracy no matter what steps you take.

Thank you for the recommendations, I will definitely increase the images quality.
Do you know if it is possible to manually select cells using QuPath?

The closest I have done is drawing the cell using the annotation tools, then converting the outline into a cell object, along with nuclear detection. It was not scalable so I never really revisited the idea.
It was mostly intended for cells on a slide/dish, so won’t work as well for tissue due to the cells overlapping. Which is a problem with most tissue based cell detection - cell borders are difficult to determine since even a thin two dimensional slice is three dimensional and will often have more than one cell within a given pixel.

CellProfiler might have some better options, if you start a topic there. But I doubt you will find anything straightforward and easy since the information you want to use (cell borders) is not included in your image.